Accessing files

Accessing files

To perform any kind of input or output (to the screen, to a file, to the network, etc), your Haskell program must use IO monad. Typically, a Haskell program begins with its main function, which has this type signature:

main :: IO ()

Because main is IO-monadic, it can call other IO-monadic functions as well as non-monadic (pure) functions. Pure functions can call other pure functions, but not IO-monadic functions (unless they cheat).

Sorting data

Ultimately, we want to sort the contents of a file that contains columns of data. For example, imagine if the file looks like this:

Jako   23    99
Eva    42    5000
Katka  0     2

Here there are three columns. We want to write a program that can read the file, sort by any of the columns, and output a sorted file.

Pure code

First, let’s write some pure functions for manipulating data.

The columnize function should take the lines of the file, as a list of strings, and break it into a list of a list of columns.


Test>
[["Jako", "23", "99"], ["Eva", "42", "5000"], ["Katka", "0", "2"]] :: [[String]]

The flexisort function will take a column number and the columnized file data, and will sort it by that column.


Test>
[["Eva", "42", "5000"], ["Jako", "23", "99"], ["Katka", "0", "2"]] :: [[String]]
Test>
[["Katka", "0", "2"], ["Jako", "23", "99"], ["Eva", "42", "5000"]] :: [[String]]
Test>
[["Katka", "0", "2"], ["Eva", "42", "5000"], ["Jako", "23", "99"]] :: [[String]]

Use the sortBy function in Data.List. Consider the sortBy (compare `on` x) paradigm.

Note that we are sorting by string representation, so in the above example, 5000 is sorted before 99, because "5000" < "99".

Doing I/O

Write an executable Haskell program (with a main function) named MySort that will do the following:

  1. Get two filenames and a column number from its command line. Use System.Environment.getArgs to get the data, and use read to convert the string to an integer.
  2. Read the content of the filename given as the first parameter (using various functions in System.IO).
  3. Sort the lines of that file, using your columnize and flexisort functions.
  4. Write the sorted lines to the filename given as the second parameter.

You can compile your program using the ghc command. After you compile it, you should be able to run it like this, assuming there is a textfile named input in the current directory:

./MySort input output 2

Write your program twice: first using Haskell’s monadic bind notation (>>= and >>), then using Haskell’s do notation.